Senior AI Engineer

About Brellium Brellium’s mission is a big one – to improve the standard of care across the US healthcare system. We’ve built AI-powered technology that helps healthcare providers deliver safer, higher-quality care - starting with the first real-time medical review platform built to fix clinical and compliance risks before they impact patients.

Each year, 1 in 20 people in the U.S. experiences a medical diagnostic or compliance-related mistake. Most providers lack the time, staffing, and tools to mitigate these issues - so they go unnoticed, impacting care quality and increasing clinical and financial risk.

Brellium is building the AI-powered platform that helps providers deliver safer, more consistent care by mitigating risk early and aligning patient visits with clinical best practices. Our goal is to give every provider in the U.S. the tools to deliver clinically excellent, data-driven care - at scale.

Brellium was founded in 2021. Since then, we’ve grown to serve over 250,000 providers across all 50 states who use Brellium to take better care of their patients and ensure data-driven, compliant care. We’re a Series A company with over $30MM in funding from First Round Capital, Left Lane Capital, and Menlo Ventures.

THE ROLE We’re hiring a Senior AI Engineer to own and evolve the intelligence engine powering Brellium. This role goes well beyond toy demos — it requires deep fluency in prompt design, evaluation frameworks, agentic systems, structured extraction, tool use, retrieval pipelines, and performance tradeoffs. The work is core: transforming raw session data into clinically structured outputs, with a constant focus on improving accuracy, latency, reliability, and cost efficiency at scale.

What You’ll Do

• Design and improve LLM-powered pipelines and agent architectures
• Build robust evaluation and benchmarking frameworks to measure output quality and regression
• Optimize for latency, cost, determinism, and output quality across high daily volume
• Implement retrieval-augmented generation (RAG) and tool-use strategies where appropriate
• Ship production-grade AI systems – not experiments
• Collaborate closely with backend and product teams to integrate AI capabilities into the core platform

YOU MIGHT BE A FIT IF YOU

• Have built and deployed LLM systems in production at meaningful scale
• Understand modern agent patterns: tool calling, RAG, memory, orchestration
• Think deeply about evaluation, hallucination mitigation, and reliability
• Care about performance and operational excellence, not just capability
• Are excited to build a core engine that operates at massive daily volume in a high-stakes clinical domain

Required Skills

• 5+ years of professional software engineering experience
• Hands-on experience building and shipping LLM-powered applications in production
• Deep familiarity with LLM concepts: prompt engineering, structured extraction, function/tool calling, RAG, evaluation frameworks
• Experience with AWS or comparable cloud infrastructure for AI workloads
• Strong software engineering fundamentals – you write clean, maintainable systems, not just notebooks

Compensation: $230,000 – $300,000 We are committed to offering a comprehensive and competitive total rewards package, including robust health benefits, commuter benefits, and meaningful ownership opportunities through equity. Compensation decisions are made holistically, ensuring fairness and alignment with market benchmarks while recognizing individual contributions and potential. Benefits Offered Include

• 401(k) Retirement Savings Plan
• Equity Compensation
• Dinner Provided via DoorDash & stocked kitchen for NY employees
• Medical, Dental, and Vision coverage coverage of up to 100% premiums for you and your family
• HSA / FSA
• 11 paid holidays each year
• Unlimited PTO
• Training and professional development
• Hybrid Work Schedule (4 days onsite, 3 if located > 1 hour away)

What it means to be “One of Us” Bias to Action: Brellium teammates do not wait to make reversible decisions or seek unnecessary approval. We quickly decide and move forward. If the decision was incorrect, we quickly reverse it and move forward.

Thinks for themselves: Brellium teammates do not take things at face value. We ask “why” until base truth is reached. If a better solution is present, Brellium teammates use it, regardless of status quo.

Negative Maintenance: The opposi…